BriefStrategyInnovation

Why Big-Bang Migrations Fail

Speed is not the risk in leaving the core. Moving everything at once is.

OMBIRO1 min read
EVERYTHING MOVED AT ONCEEvery product, every customer, one cutover: the risk multiplies
£48.65mcombined FCA and PRA fines on TSB, December 2022

In April 2018 TSB moved all of its customers to a new platform at once. The data moved; the platform failed. Branch, phone, online and mobile banking were disrupted, a significant share of 5.2 million customers were affected, and the bank was not back to business as usual until December.

Commonwealth Bank of Australia shows the slower version of the same problem: a core programme budgeted at A$580m in 2008 ended near A$1.1bn.

Neither case argues for staying on the old core. Both argue for moving product by product, running old and new side by side, and switching the suite off only when nothing depends on it.

Exhibit 5Contain, hollow out, retire
HORIZON 1 · 0–12 MONTHS Contain HORIZON 2 · 12–36 MONTHS Hollow out HORIZON 3 · 36–60 MONTHS Retire and scale Legacy core Freeze customisation; wrap in APIs shrinks as products leave retired Ledger New products on acloud-native ledger Migrate product by product,parallel run, no big bang One ledger of record;suite switched off Products First product liveon an engine Deposits, cards, lendingmove to engines Every product on engines;launch in weeks Agents Staff copilots; agentgateway pilot Autonomous servicing andcollections within limits Serve customer delegatesand partner agents at scale Accountability Continuous reconciliation;mandate registry Automated regulatoryreporting; model risk office Real-time supervisionfeeds; resilience tested

The three-horizon agenda from the Big Idea on leaving the core.

Read more on Strategy